Magic Secrets Revealed

Sunday, February 11, 2007

A guide on doing keyword research for SEO

This is a long article explaining in detail about doing keyword research for SEO. An important piece for web publishers.

How to do Keyword Research for Search Engine Optimization

There are many ways to do Keyword Research but the fundamentals of how it is done is pretty much the same even if every SEO professional has their own way of doing it when looking at the very specific details. The methods discussed here may not be the best, but through SEO experimentation, this is what has worked best for me and my clients. For the purposes of discussion, when I refer to the word keyword, this may not necessarily referring to 1 word alone but to many keywords, or sometimes called a keyphrase. But if I refer to keyword alone, it may mean a group of words or a keyphrase.

The raw keyword list to start your keyword research

I simply list down the product or service I am optimizing for and come up with a good list of synonyms and related words for the keyword. Just to come up with a random keyword for example purposes, let's say the client sells cars. I will come up with the synonyms:

Car
Auto
Automobile
Vehicle
Toyota (Brand)
Celica (Model)

And may also come up with related words like:

Drive

It will be up to you if you think the word is related or not. I am sure you can come up with a very long list, but don't make it that long. It has to be something you think people may type in when searching to buy your car. You can add in more word that you think other people might add when searching for a product or service to buy.

Buy a car
Car buying
Car store
Automobile shop
Toyota dealer

If you are doing SEO to market a physical store, might as well add in the geographic location of your shop. For example purposes only, let's say the car store is in a place called Gotham City, Metropolis. Then I will considered adding in these words in my keyword research. But if you are selling something small that is shipped online and you are targeting to sell nationwide or even internationally, you may opt not to pay much attention to adding the location in the keyword research.

Coming up with synonyms and related word, there are many tools you can use. Sometimes all these tools can be overwhelming just to come up with synonyms and related words. If you are on a tight budget, and want to use something for free, Google alone can help you with that. Do a related keyword search using the tilde ( ~ ) operator. Like search: ~cars or use the Adwords Keywords Tool.

Checking current search volume for your targeted keywords

After coming up with your raw keyword list, how often do you think these keywords are searched? What are the actual variations of these keywords that were searched by other people? You have several tools you can use to guide you in knowing these data. Here are a few of the tools available:

Adwords Keywords Tool - You can determine the highly searched keywords here and also keywords that companies are paying a lot for.
Overture Keyword Suggestion Tool - This will give you the number of searches for a keyword on Yahoo for the previous month. I will get the relevant keywords listed and the search number on overture and place them on another column of the spreadsheet.
Wordtracker - One advantage of Wordtracker over just relying on Overture is Overture cannot seem to see the difference between plural and singular words. Aside from that, the keyword order in Wordtracker does not seem to matter on many occasions. Like New Cars and Cars New will be the same on Overture, but not on Wordtracker. Wordtracker is not a free tool, but it is not that expensive either. They have a free trial and if ever you just need to use it for a small campaign, you can subscribe to the service for a single day and you can do so much research in 1 day.
Keyword Discovery - Now one major advantage to use Keyword Discovery over all of the above is the data they have up to 12 month of past data.

All this data, you can tabulate on a spreadsheet, having 1 column for each taking note of the search volume. If you have a physical location, a store you are promoting, then this is where I try the keywords with the location in it. For our example, I would do: cars gotham metropolis.

All these tools are good but do not treat them like they are the absolute truth. Please note the following:

To get a good comparison of results, take note of the unit of time you have on the data. If it is per month or per year and keep units the same by doing some conversions.
Some keywords will be present in all Google, Overture, Wordtracker and Keyword Discovery. Prioritize the keywords that appear in all of them and give less priority to those that are only in one of the keyword research tools.
Looking at all tools gives you the big picture, but the results of all the tools above can give you varied results. By experience, most of the time the highest result is on Overture, next is Wordtracker and the smallest values are on Keyword Discovery. Please note that all these depend on the sources. Overture is from Yahoo, Wordtracker is from Dogpile and Metacrawler. Keyword Discovery is from a large number of various engines. Also a common trend I see is for brand name search, the highest will appear on Overture, and my guess is it's caused by the client themselves that keep searching for themselves.

Check how much competition do you have for your targeted keywords

Now that you have several keywords gathered using the tools above, how much competition do you have for every keyword? Searching for the keywords on Google, Yahoo and MSN will give you a total number of search results. Now if you see 1,000,000 or more keywords, does that mean you are competing against 1,000,000 websites for the number 1 spot? These search engines are pretty powerful and do remember that they will search every website within their database where all occurrences of each keyword are on a webpage will be part of the results. So all the total results may be the number of websites you have to outrank but not every one may really be targeting the keywords you are targeting too. Now in all of the results listed, how many are actually from a competitor that is optimizing their website and how many are just there because their site has the words you were targeting?

Since part of SEO is placing keywords in the URL, like in domains, subdomains or folder names. In the title tags and in anchor text in internal links and external linking. You can pretty much search on these using several search commands like, intitle, inurl and inanchor. For our example, we can search:

intitle:"cars gotham metropolis"

Notice that I placed the keywords in quotes since most people search:

intitle:cars gotham metropolis

This will search for the keywords but it is only requiring the word cars in the title tag. Searching:

intitle:cars intitle:gotham intitle:metropolis

Will give you results that have all 3 words in the title tag. But having them in quotes, will look for the exact same phrase in the title tag. Thus you get to narrow down the results to really how many websites are trying to compete for your same keyword.

Which keywords do you target from your researched keyword list

Let's say you have this situation: Keyword option 1 was searched 1,000 times the previous month but has 100,000 in competition actually targeting the same keywords. While keyword option 2 was searched less at 500 times a the previous month but has 900 in competition targeting the same keywords. Even if option 1 was searched more the previous month, it might be more wise to concentrate on option 2 since you have such a low amount of competition and getting higher in the SERPs will be a lot more easier.

For keyword optimization, you want the highly searched but low search competition. Many people observe this ratio in the KEI or Keyword Effective Index. This is basically the number of times a keyword is searched over the number of competition. If you wish to multiply this by 100 to get it in terms of a percent, you may do so. Everyone may have their own formula for it.

How I would personally do it is I start with keywords that were found in all these four: Google, Overture, Wordtracker and Keyword Discovery as my priority keywords. I will give less priority for keywords that only appear in one or two of them. Add the times they were searched in all 4 and divided by the total competition on Google, Yahoo and MSN.

(Overture + Wordtracker + Keyword Discovery) / (Intitle results on Google, Yahoo and MSN) * 100 = KEI

Some SEO professionals may have a different KEI formula but there is really no correct formula. As long as you get a value where you can gauge what keywords are highly searched but with low competition. The highest KEI value will be the main keywords to target.

Now let's say you got the top 50 keywords based on KEI, do you target all of them on all of your pages?

How many keywords to target per page

Several SEO's may have different beliefs on this, but based on SEO experiment experience, I target only 2 to 3 keywords per page and focus on those alone. Then the rest of your keywords can be targeted on other pages of your website. Spread them out where you select the most appropriate page where they will be focused on. Targeting too many keywords on a single page just brings dilution of the true meaning of the page.

How long of a tail should a long tail be?

So what is this long tail? You hear it in SEO, you hear it in PPC, although they may not exactly be used in the exact same way. Here are the general stuff first:

The lesser the keywords, the more general it is, the more competition you have, that harder it is to rank high in the search engines. But these terms would also probably have a high searched volume.
The more keywords you have, the more specific the search term is, the lesser competition you will have, the easier it gets to rank higher in the search engines. This would probably have a lesser searched volume.

Now general known behavior about online buyers:

People that search general terms are usually people doing some research about a particular topic. Some may be buyers as well and as they get tons of results, they might narrow down their search by being more specific.
People that search specific multiple keyword terms really know what they are looking for or may have already narrowed down their search into something more specific to really find what they wanted to buy. Usually the specific term searcher has more potential to buy than the general term searcher.

Hitting two birds with one stone by using the long tail.

With a list of keywords gathered based on you research using the tools above, many keywords with high KEI could be targeted all at the same time in some cases. Let's say in our case, we came up with these keywords:

1. cars
2. cars for sale
3. car buying
4. buying cars in gotham city
5. buying cars
6. cars for sale in gotham city metropolis
7. car buying in gotham

In this case, we can target number 6, the long tail "cars for sale in gotham city metropolis" and at the same time you get to target number 1 and 2 as well since they are part of number 6. Number 7, "Car buying in gotham" will target number 1 and 3. Number 4 "Buying cars in gotham city" will target number 1 and 5.

To sum it all up, keyword research is:

1. Coming up with your raw keyword list.
2. Check the KEI using tools to check the past search volume over the number of competition.
3. Get the long tails and group together with short general terms that can be targeted at the same time.
4. Assign about 2 to 3 keywords to a single page to focus on and distribute among pages and give instructions to your SEO writer.

The rest of the SEO will be on writing content and building the site with the proper code which are all part of onpage SEO. When the whole site is launched, offpage SEO will follow.

Article source: Benj Arriola (ArticleBiz.com)